Soils throughout the world represent a carbon reservoir of 615 billion tonnes in the layer down to a depth of 20 cm, and 2344 billion tonnes if this inventory is extended to a depth of 3 metres.  Taken together, these stocks represent more than biomass and atmospheric CO2 combined.  The mean residence time for organic carbon in the soil increases markedly with depth, reaching between 2000 and 10,000 years in so-called deep layers, i.e. below 20 cm.
However, although the mechanisms underlying carbon decomposition by the action of soil micro-organisms are known, the factors governing the stability of organic carbon in deep soil layers have not been precisely identified.  Knowledge of these factors is thus essential to determining whether climate change will or will not activate decomposition of this major, deep carbon reservoir, and thus accelerate the increase in atmospheric CO2. 
To identify these parameters, INRA scientists have used a soil profile situated in the Environmental Research Observatory (ORE) on permanent grassland set up in the Massif Central region by the INRA Clermont-Ferrand Research Centre in 2003.  This site has been occupied by permanent grassland for the past 50 years, and 2000 years ago was covered by hornbeam and chestnut forests.  Carbon 14 dating of the soil carbon indicates that the organic matter stored in deep layers are ancient and arose from these forests that have now disappeared.
The researchers studied a deep layer of this soil (between 60 cm and 80 cm) containing organic carbon with a mean age of 2600 years.  They observed that microbial degradation of this ancient carbon was markedly stimulated when plant litter was added to it.  This litter, which is not present in a natural state in these deep layers, provided the energy necessary for the development of the micro-organisms that decompose "humified" carbon in the soil.
This result confirms an idea already put forward by the same team in 2005 that it is the absence of plant litter – an essential source of energy for soil micro-organisms – that is crucial to the stability of organic carbon in deep soil layers.

These results suggest that, even if conditions of temperature and humidity are favourable for microbial activity, the organic carbon present in deep soil layers does not contain sufficient available energy to feed active microbial populations so that they will produce the enzymes necessary to decompose carbon.  The existence of this energy barrier could thus reduce the effects of global warming on the decomposition of these considerable deep carbon reserves, which is in contradiction with the effects forecast by current models, based as they are on an acceleration of the rate of enzymatic reactions in line with rising temperatures (Arrhenius' theory).
Storing carbon in deep soil layers (where it remains stable in the long term) thus constitutes an interesting alternative to current approaches that are based on the short-term storage of carbon in the form of plant biomass or organic carbon in surface soil layers.

The results also show that the decomposition of deep organic carbon can be reactivated, as was achieved by the research team.  This means that changes to land use and agricultural practices that encourage the penetration of roots and the deposition of litter at low levels (e.g. deep soil tillage or the use of drought-resistant plants with deep roots) could stimulate the loss of this ancient carbon.
Further studies in other soil types will be needed to supplement these findings with a view to validating them at a larger scale.  However, given that microbial decomposition is limited by the availability of energy in numerous soil types, and that the low availability of litter at depth is based on a fundamental property of ecosystems (plants live and deposit their litter on the surface), the scientists consider that these results could be generalised to numerous soils (with the exception of flooded or frozen soils) in which microbial decomposition is limited by other factors.